Topics in Functional and Ecological Vertebrate Morphology, pp. 205-236. P. Aerts, K. D’Août, A. Herrel & R. Van Damme, Eds. © Shaker Publishing 2002, ISBN 90-423-0204-6

Evolutionary Relationships between Cranial Shape and Diet in Bats (Mammalia: Chiroptera)

Victor Van Cakenberghe 1*, Anthony Herrel 1 & Luis F. Aguirre 2

1 Department of Biology, University of Antwerp, Belgium 2 Centro de Biodiversidad y Genetica, Universidad Mayor San Simon, Cochabamba, Bolivia

Abstract

The trophic radiation of new world phyllostomid bats has often been cited as an example of an explosive adaptive radiation. However, among Old World bats a similar radiation into diverse feeding niches such as frugivory, carnivory, insectivory and nectarivory has occurred. Previous analyses of cranial shape in dietary specialists have indicated general trends in cranial shape that seemed to be correlated with differences in diet. However, up till now, a comprehensive analysis of correlations between cranial shape and diet in an explicit phylogenetic context has never been performed. Here we analyse data for 140 species of bats from 11 families, comprising most of the major radiations. Principal component analyses coupled with analyses of variance indicate several variables that seem to be informative in explaining variation among different trophic morphs. Comparative analyses, taking into account the phylogenetic relationships among the species studied, indicate that evolutionary shifts in diet among bats are indeed correlated to changes in cranial shape. Changes in cranial shape were largely in accordance with a priori predictions based on functional demands imposed by different diets. However, in contrast to our predictions, no differences were observed between bats specialising on hard versus soft prey. The absence of such differences is most likely the result of a general lack of 1) quantitative dietary information for many species and 2) the absence of data describing functional properties of food items.

Key words: Chiroptera, bats, morphometrics, diet, phylogenetic analysis, ecomorphology.

Introduction

The trophic radiation of New World phyllostomid bats has been cited as example of an explosive adaptive radiation (Marshall, 1983, Freeman, 2000). Phyllostomids have diverged into highly specialised feeding niches such as nectarivory, sanguinivory, carnivory and frugivory from a presumably insectivorous ancestor. However, a similar radiation into these feeding niches (with the exception of sanguinivory) has also arisen among Old World bats. Yet, in contrast to New World phyllostomids, no single family of Old World bats has radiated that extensively. Both within Old and New World bats, the radiation into novel trophic niches has been accompanied by extensive morphological specialisation (e.g. Freeman, 1979, 1984, 1995) to match the changes in functional demands imposed by different food types on the cranial system. Previous analyses of cranial shape in dietary specialists have indicated general trends in cranial shape that appear to be correlated with differences in diet. Freeman's ground-breaking work

* Address for correspondence: Dr.Victor Van Cakenberghe, Department of Biology, University of Antwerp (UIA), Universiteitsplein 1, B-2610 Antwerpen (Wilrijk), Belgium. E-mail: [email protected] 206 Van Cakenberghe et al. showed that bats specialising on hard-shelled prey ( e.g. beetles) have thicker jaws, better developed cranial crests and changes in dental structure compared to closely related species eating mostly softer insects (Freeman, 1979). Frugivores specialising on hard fruits (e.g. figs) show behavioural modifications that presumably allow them to maximise their bite forces (Dumont, 1999). Thus, hard fruit specialists are also expected to have ticker jaws, better developed cranial crests, and more blade like teeth. Freeman (1984) also demonstrated how carnivorous bats tended to have more elongated skulls, larger brain volumes, and dental modifications which enhance grasping ability. Nectar specialists, on the other hand have long and narrow heads (Freeman, 1995; Nicolay and Dumont, 2000). However, megachiropteran nectarivorous bats tend to differ from other nectarivores in the width of the head, supposedly allowing for more jaw musculature. As this seems functionally not correlated with nectarivory, it has been suggested that this might be a remnant of the ancestral frugivorous feeding strategy of this group. Despite these general trends, a recent study on resource partitioning in phyllostomid bats revealed few or no correlates between bite performance and some simple external head shape variables (Aguirre et al., 2002). Up till now, this relationship has always been assumed to underlie the observed differences in cranial structure among different trophic morphs (e.g. Freeman, 1979, 1984). The absence of a correlation between head shape and bite performance might be simply the result of a lack of resolution in the head shape variables measured by Aguirre et al. (2002). The observed variation in head shape among feeding specialists (e.g. Freeman, 1979, 1984, 1995), however, might not be adaptive, but rather the result of historical phenomena, or co-evolutionary changes in response to selection on other traits (e.g. traits related to flight, echolocation, etc…). By taking into account the historical relationships among different species, and by quantitatively analysing the variation in cranial shape among trophic radiations within all major radiations of bats, such non-adaptive scenarios can be detected. A recent analysis of the relationships between cranial morphology and diet among ungulates (Perez-Barberia and Gordon, 1999) demonstrated the importance of incorporating the evolutionary history of the species under study. Of the 15 traits that were significantly different in a traditional analysis only two remained significant after taking into account the relationships between the species. Still, distinct differences between ungulates specialising on tough plant matter versus non-specialist species could be demonstrated based on differences in the size of the coronoid process and the height of the skull. Therefore, we decided to quantitatively examine skull shape in an explicit phylogenetic context for 140 species of bats including all major radiations (see appendix 1, 2). In our analysis, we test the following a priori predictions, based on previous research (e.g. Freeman, 1979, 1984, 1995, 2000) and biomechanical considerations. We predict that bats specialising on hard prey (either insects, fruits or vertebrates) will have generally more robust skulls (i.e. wider and higher) than other bats (Freeman, 1979, 1981). Additionally, we predict longer lower jaws (Freeman, 1979, Barlow et al., 1997), higher coronoid processes (Jacobs, 1996) and longer distances from the condyle to the coronoid process (effectively increasing the jaw closing in-lever). We also predict hard object frugivores to have higher coronoid processes than hard object insectivores, as these animals will typically be biting with large gapes (note that high coronoid processes may be advantageous for two reasons: 1) increased surface area for the attachment of the jaw muscles; 2) to optimise the orientation of the temporalis muscle at large gape angles). Carnivores should have high coronoids and skulls (allowing for the attachment of well-developed jaw muscles; see Freeman, 1984), which should be robust to counter the reaction forces generated while biting. Moreover, carnivores should have high coronoid processes and relatively low condyles (i.e. low jaw joints relative to the toothrow) to optimise bite performance at large gapes (Turnbull, 1970). Nectarivores are predicted to have long narrow snouts (which should be reflected in the relation between the length of the upper toothrow and the interorbital width, the width across the upper molars and the width across Cranial Shape and Diet in Bats 207 the canines) allowing them to penetrate corolla of flowers (Nicolay and Dumont, 2000), and to support the long and mobile tongue (Freeman, 1995). As the demand for bite strength is likely reduced (see also Nicolay and Dumont, 2000; Aguirre et al., 2002) we expect a reduction in skull height and width, and in the size of the coronoid process. For sanguinivores, we expect a moderately long snout to support the tongue, which is used to lap up blood. Due to its highly specialised dentition (which is used to cut rather than to chew), we also expect a decreased importance of the masticatory system (Aguirre et al., 2002) which would be reflected in relatively narrow and low skulls.

Materials and Methods

Specimens We examined the cranial measurements of more than 18.000 bat specimens representing most of the Chiropteran families. Details on the specimens examined can be found in Claessen and De Vree (1990a, 1990b) for Epomophorus, De Vree (1969) for Mops spurrellii, Van Cakenberghe and De Vree (1985, 1993a, 1993b, 1999) for Nycteris, Robbins et al. (1984) for Scotophilus, and Van Cakenberghe and De Vree (1994) for Rhinopoma. The data for the species in the Pipistrellus/Eptesicus genus complex were obtained during the senior authors' current taxonomic study into the systematics of this complex. Additionally, measurements from specimens collected during field trips to Côte d'Ivoire (De Vree, 1971), Togo (De Vree et al., 1969, 1970), Rwanda (Baeten et al., 1984) and Zaire (Van Cakenberghe et al., 2000) were included. Data for additional species of the families Phyllostomidae and Pteropodidae were gathered recently during visits to the Koninklijk Belgisch Instituut voor Natuurwetenschappen, Brussels (Belgium) and Naturalis, Nationaal Natuurhistorisch Museum (former Rijksmuseum voor Natuurlijke History), Leiden (the Netherlands). In the present study, we followed the taxonomy presented by Koopman (1993), with modifications according to the above mentioned publications. Additionally, we followed Harrison and Bates (1991), Corbet and Hill (1992), Peterson et al. (1995), Horácek et al. (2000), and Hutson et al. (2001), and consider the Hipposideridae a valid family. The taxonomy for the Phyllostomidae follows Simmons and Voss (1998) and Wetterer et al. (2000). For the Vespertilionidae, the most progressive subdivision for the Pipistrellus/Eptesicus complex is followed which include the recognition of Arielulus, Eptesicus, Falsistrellus, Hypsugo, Neoromicia, Perimyotis, Pipistrellus, and Vespadelus as separate genera. For the analyses, a subset of data was retained from the original data set based on the following criteria: whenever data were available for several species within a group lacking in resolution in its historical relationships (i.e. polytomies) and without variation in ecomorph type, the species with the largest sample was chosen to represent that group. Additionally, species which lacked more than four of the cranial characters included were eliminated from the analysis. This resulted in a data set comprising over 8.000 specimens of 140 species of 11 families of bats. Species averages were calculated for all characters and used in subsequent analyses.

Morphometrics All cranial measurements were taken using digital or vernier callipers and were rounded to the nearest 0.1 mm. Measurements included (Fig. 1): the greatest length of skull from the base of the canines (Gslc); the condylo-canine length (Cblc); the least interorbital width (Ior); the width across the zygomatic arches (Zyg); the width of the skull across the mastoid processes (Mast); the width of the braincase (Brain); the length of the upper (maxillary) toothrow (c-m3); the width across the upper canines (c-c); the width across the posterior upper molars (m-m); the height of the skull over 208 Van Cakenberghe et al. the bullae (Shpb); the length of the mandible (Mand); the length of the lower (mandibular) toothrow (c-m3); the length of the underside of the mandible (from the symphyse between the rami to the angular process, Mand2); the distance between the condyle and coronoid processes (ConCor); the height of the coronoid process (Cor); the distance between the condyle and angular process (Conang) and the forearm length (Fa).

Figure 1. Cranial measurements taken on the species in the analysis. Gslc: greatest length of skull from the base of the canines; Cblc: the condylo-canine length; Ior: least interorbital width; Zyg: width across the zygomatic arches; Mast: width of the skull across the mastoid processes; Brain: width of the braincase; c-m3: the length of the upper (maxillary) toothrow; c-c: width across the upper canines; m-m: width across the posterior upper molars;

Shpb: height of the skull over the bullae; Mand: length of the mandible; c-m3: the length of the lower (mandibular) toothrow; Mand2: the length of the underside of the mandible (from the symphyse between the rami to the angular process; ConCor: the distance between the condyle and coronoid processes; Cor: the height of the coronoid process; Conang: the distance between the condyle and angular process.

Diet Dietary data for all species were compiled from literature data (Appendix 1). Based on these data, bats were categorised as belonging to one of eight dietary categories: 1: Sanguinivore (blood lapping species), 2: Nectarivore (nectar eaters, also including palynivore or pollen eating species), 3: Carnivore (species eating vertebrates), 4: soft insect eaters (eating soft arthropods such as moths, crickets, spiders), 5: hard insect eaters (species that include hard arthropod prey such as beetles or scorpions into their diet), 6: soft fruit eaters (species specialising on soft fruits such as bananas, Piper, Solanum), 7: hard fruit eaters (species specialising on figs, or woody fruits), and 8: omnivores (species known to consume fruits, insects, nectar and vertebrates). We are aware that this classification does not represent the true variation in diet within each species. However, classifying Cranial Shape and Diet in Bats 209 bats into functionally relevant dietary groups allows us to test previously proposed hypotheses concerning the relationships between cranial morphology and diet in bats. Based on the results of the initial principal component analyses, we lumped all insectivores into a single 'insectivorous' category for further analysis. Additonally, hard and soft object frugivores were lumped into a single 'frugivorous' category, and omnivores were grouped with the carnivores into a single category.

Non-phylogenetic analysis Species means were calculated for each cranial variable and entered into a principal component analysis. As the purpose of this analysis was to explore relations between cranial shape and diet only species for which a full data set was available were included in this analysis (species indicated by an asterisk in appendix 1). This data set included representatives of all trophic groups and all families examined. We performed a principal component analysis with varimax rotation, and we extracted the first four components for further analysis. We used the scores of each species on the first four principal components in a MANOVA to test whether trophic morphs differed in cranial size and shape. Next, we applied univariate ANOVA's on the scores on each of the four components to test which of these explained variations between the trophic morphs. Finally, we performed post-hoc tests (Tukey HSD) to test, which of the trophic morphs differed from one another on the first four axes.

Phylogenetic analysis Based on the results of the non-phylogenetic analysis, five variables proved to be informative in explaining variation between trophic morphs (see tables 2,3). For these five variables, we calculated species means. For two of these (c-c, Mast), data were available for all 140 species in the analysis. For m-m, data were lacking only for Lavia frons, and for Ior, data were lacking only for the Nycteridae. For Shpb, data were available for all species indicated with an asterisk in appendix 1. As species share part of their evolutionary history, they cannot be considered independent data points (Felsenstein, 1985; Harvey and Pagel, 1991). Hence, the statistical significance of differences among sets of species (in this case among trophic morphs) cannot be evaluated with standard F-distributions. To overcome this, appropriate empirical null distributions, taking the phylogenetic relationships among species into account, can be created using Monte Carlo simulations. Here, we used the PDSIMUL and PDANOVA programs developed by Garland et al. (1993). The Brownian motion model is used as model for evolutionary change and we ran 1000 unbounded simulations. Next, we used the PDANOVA program to generate a new F-value based on the simulated data sets. Differences among trophic morphs were considered significant if the traditional F-value (F95) was larger than the F-value calculated based on the simulated data sets (Fphyl). As these methods require information on the topology and branch lengths of the phylogenetic tree linking the species under investigation, we constructed phylogenetic trees (see appendix 2) based on the super tree for bats from Jones et al. (2002). Different trees were constructed for the different species groups in the analysis (depending on the availability of cranial measures, see higher). As little information is available on divergence times between species, all branches were set to unity (see Diaz-Uriarte and Garland, 1998 for the validity of this approach). In the analysis, we tested all variables separately, and we adjusted significance levels using sequential Bonferroni corrections (Rice, 1989). 210 Van Cakenberghe et al.

Results

Non-phylogenetic analysis The first four components of the principal components analysis (PCA) explained over 98% of the variation in the data (Table 1). We used the scores on the first four principle components (PCs) to perform a multivariate analysis of variance, which indicated significant differences between trophic categories (Wilk's Lambda: 11.373; P < 0.001). Subsequently, we carried out a one-way analysis of variance, which revealed significant differences on all principal components. The first principal component correlated strongly with most of the cranial variables included in this analysis. This component should be considered as an overall indicator of size, as large species such as Pteropus vampyrus (Gslc: 69.8 to 82.9 mm) and most of the other large Old World fruit bats showed high scores on this axis (Fig. 2). Differences between dietary categories were no longer significant on the first principal component after Bonferroni correction, which indicates that trophic morphs did not differ in overall skull size (F7,87 = 2.14; P = 0.047). Additionally, Post-Hoc tests performed on the scores on the first axis indicated no significant differences between dietary groups (all P > 0.15).

Variable PC1 PC2 PC3 PC4

% variance 45.40 19.74 19.26 13.92 eigenvalue 7.26 3.16 3.08 2.23

Gslc 0.730 0.372 0.461 0.324 Cblc 0.768 0.390 0.393 0.311 Ior 0.451 0.747 0.375 0.283 Zyg 0.683 0.483 0.402 0.353 Mast 0.450 0.437 0.628 0.447 Brain 0.648 0.490 0.467 0.334 c-m3 0.747 0.299 0.377 0.441 c-c 0.527 0.550 0.382 0.487 m-m 0.493 0.460 0.422 0.597 Shpb 0.531 0.421 0.681 0.259 Mand 0.785 0.400 0.347 0.312

c-m3 0.756 0.332 0.358 0.422 Mand2 0.761 0.394 0.359 0.348 Concor 0.683 0.378 0.489 0.353 Cor 0.819 0.446 0.284 0.176 Conang 0.766 0.314 0.414 0.317

Table 1. Loadings of variables on first four principal components after varimax rotation. (Bold values indicate high loadings on that axis. Note how most variables strongly correlate with the first principal component.)

The other components extracted primarily explained variation in skull shape rather than size. The second principal component had high positive loadings with the least interorbital width (Ior) and the width of the skull as measured across the canines (c-c). One way analysis of variance indicated significant differences between trophic morphs along this axis, even after Bonferroni correction

(F7,87 = 6.87; P < 0.001). Post-Hoc tests indicated significant differences between both hard and soft fruit eaters and the other dietary categories (with the exception of sanguinivores and omnivores; see Fig. 2) on this axis. Cranial Shape and Diet in Bats 211

Figure 2. A) Position of the different trophic morphs in morphospace as described by the first two principal components. The first principal component is cleary a 'size' axis with large species such as Pteropus vampyrus showing high scores on this axis. The second axis indicates trends in cranial shape that are significantly related to variation in diet. B) Position of the different trophic morphs in cranial morphospace as described by principal components 2 and 3. Whereas the second principal component is positively correlated with the least interorbital width (Ior) and the width across the canines (c-c), the third principal component is positively correlated with skull height (Shpb) and the postorbital width of the skull (Mast). The frugivorous bat, Hypsignathus monstrosus has a highly derived cranial shape as indicated by its high score on the second principal component. However, this variation is unrelated to its feeding strategy. Both the carnivores and the sanguinivores clearly stand out in the cranial space described by these two axes. Closed circles: soft object frugivores; open circles: hard object frugivores; closed squares: soft object insectivores; open squares: hard object insectivores; diamonds: nectarivores; hexagons: carnivores; upward facing triangles: sanguinivores; downward facing triangles: omnivores. 212 Van Cakenberghe et al.

Exploration of the means for each dietary category (Table 2) indicates that this is because frugivores have wider interorbital distances and wider snouts at the level of the canines.

Trophic Inter-orbital Post-orbital Width across Width across Skull height morph width (Ior) width (Mast) canines (c-c) posterior upper (Shpb) molars (m-m)

Sanguinivore 5.69 ± 0.44 12.74 ± 0.37 5.63 ± 0.33 5.79 ± 0.69 13.41 ± 0.55 Nectarivore 4.95 ± 1.34 10.30 ± 1.73 4.90 ± 1.49 6.38 ± 1.11 9.26 ± 1.57 Carnivore 6.17 ± 1.38 15.88 ± 3.62 7.47 ± 1.47 11.47 ± 2.33 14.17 ± 2.22 Insectivore 3.96 ± 0.94 9.29 ± 2.76 5.00 ± 1.45 7.09 ± 2.00 6.77 ± 1.87 Frugivore 7.08 ± 1.89 15.00 ± 4.18 7.87 ± 2.68 11.97 ± 3.83 13.11 ± 4.36

Table 2. Averages of the skull measurements extracted from the principal component analysis for each dietary category (Table entries are averages ± standard deviations.)

The third PC correlated highly with skull height (Shpb) and the width of the skull across the mastoid processes (Mast). A univariate ANOVA indicated highly significant differences among diet groups (F7,87 = 21.67; P < 0.001). Post-Hoc tests indicated significant differences between both sanguinivores and carnivores and all other trophic morphs (with the exception of omnivores). Additionally, insectivores differed significantly from soft object frugivores and omnivores (Fig. 3). Exploration of the means for each dietary category indicates that these differences are the result of high skulls in the carnivores and sanguinivores and rather low and especially narrow skulls in the insectivores (Table 3).

Variable Fphyl F P

Width across canines (c-c) 25.70 16.44 0.132 Post-orbital skull width (Mast) 24.10 23.70 0.052 Width across maxillary teeth (m-m) 25.10 25.72 0.047 Inter-orbital width (Ior) 26.50 36.21 0.002* Skull height (Shpb) 24.80 33.38 0.019

Table 3. Results of simulation analysis testing for differences between diet groups. (* significant after sequential Bonferroni correction at the 0.05 level.)

The fourth PC correlated highly with the width of the snout as measured across the posterior maxillary teeth (m-m). A one-way ANOVA revealed highly significant differences among diet groups (F7,87 = 11.48; P < 0.001). Post-Hoc tests indicate that nectarivores and sanguinivores differ significantly from all other dietary categories on this axis, both having extremely narrow skulls at the posterior upper tooth row (Fig. 3). Cranial Shape and Diet in Bats 213

Figure 3. A) Position of the different trophic morphs in morphospace as described by the second and fourth principal components. The second principal component is positively correlated with the least interorbital width (Ior) and the width across the canines (c-c). The fourth principal component is correlated with the width of the skull across the posterior upper molars. On these two axes, nectarivores and sanguinivores are clearly separated from the rest of the trophic morphs. B) Position of the different trophic morphs in cranial morphospace as described by principal components 3 and 4. The third principal component is positively correlated with skull height (Shpb) and the postorbital width of the skull (Mast). The fourth principal component is correlated with the width of the skull across the posterior upper molars. In the space described by these two axes differences between carnivores, nectarivores, sanguinivores and the other trophic morphs are clear. Closed circles: soft object frugivores; open circles: hard object frugivores; closed squares: soft object insectivores; open squares: hard object insectivores; diamonds: nectarivores; hexagons: carnivores; upward facing triangles: sanguinivores; downward facing triangles: omnivores. 214 Van Cakenberghe et al.

Phylogenetic analysis Simulation analysis on the above mentioned variables (see Table 2) indicated significant differences in skull shape among dietary groups. However, the width of the skull (as measured across the canines) was not longer significantly different among trophic morphs (Table 3). Other variables approached significance (post-orbital skull width) or were significantly different among dietary groups. It should be noted, however, that only differences in inter-orbital skull width remained significant after Bonferroni correction. Overall, this analysis indicates that evolutionary shifts in trophic niche are indeed accompanied by evolutionary changes in skull shape.

Discussion

Of the 16 cranial measurements examined, only five were informative in explaining differences among feeding strategies in bats. Surprisingly most of these were width measurements, three of which are indicative of the shape of the rostral part of the skull. Of the other measurements taken, only skull height seemed to explain some of the variation among trophic morphs. In contrast to our a priori predictions, none of the lower jaw measurements explained variation in cranial shape among the diet categories. In addition, rather unexpectedly, none of the variables examined here distinguished between hard and soft object specialists. Both within insectivores, and frugivores, the analyses were unable to detect differences between hard and soft object specialists. Yet, previous analyses of cranial shape in molossid and vespertilionid bats indicated distinct differences between soft and hard insect specialists (Freeman, 1979, 1981; Whitaker, 1994). Although it might be argued that we did not use the appropriate variables to detect these differences, we did include measurements that were previously suggested as being different between hard and soft object specialists (e.g. the height of the coronoid, length of the mandible; see Freeman, 1979; Jacobs, 1996; Barlow et al., 1997). Alternatively, we would like to suggest the following hypotheses to explain the lack of differences between hard and soft object specialists: 1) there are no differences in cranial shape among the various hard and soft object specialists, 2) the resolution in the available ecological data influencing the classification of bats into hard and soft object specialists is insufficient, and 3) similar selective pressures result in different adaptive responses in the different groups investigated here, such that no overall response was detected (see Harvey and Pagel, 1991). Based on biomechanical theory we would like to argue in favour of the second alternative. For example, when considering vespertilionid bats, hardly any dietary information was available for more than two thirds of the species in the analyses, resulting in their classification as being 'soft' insect specialists. However, among the frugivores (especially among the phyllostomids) more resolution in diet was available. Still, no differentiation was apparent between soft and hard object specialists. This trend indicates that few or no differences in cranial shape are present. Potentially, changes in the architecture of the jaw closing system (e.g. changes in pennation angle, fibre lengths, subtle changes in the line of action of the adductors) that enhance bite performance might be achieved with only minimal changes in skull shape. Alternatively, behavioural adjustments in food processing strategies that maximise bite strength may exist. As has been observed for phyllostomids, biting with the posterior teeth minimises the jaw out-lever and can thus optimise bite capacity (Dumont, 1999). Behaviour might in this case act as a filter between morphology and performance thus blurring the relationships between morphology and ecology. More detailed biomechanical analysis of the jaw system (including a functional investigation of the jaw musculature) and behavioural observations of feeding behaviour in other hard versus soft food specialists will be especially insightful in explaining the patterns (or lack thereof) observed here. Cranial Shape and Diet in Bats 215

Based on the characters used in the analyses, omnivores could not be differentiated from carnivores. As all omnivores in the analyses (species of the genus Phyllostomus) include vertebrates into their diet, this indicates that constraints on this feeding mode are potentially stronger than for other dietary modes (i.e. that a carnivore can still be proficient at exploiting insects and fruit, but not vice versa). Indeed, carnivores were typically very distinct from other trophic morphs and had the greatest skull height. Although not significant when taking into account the relationships among species (Table 3), carnivores also tended to have extremely wide skulls, which was in accordance with our predictions. Presumably, the wide and high skulls of carnivores provide a greater attachment area for the jaw adductors. Given that the coronoid process was not markedly different from that in other trophic morphs, the higher skulls presumably allow for a more optimal orientation of the temporalis muscle. In contrast to previous studies (Freeman, 1984), the carnivores included in the data set did not have longer skulls than representatives from other trophic groups. The nectar specialists in the analyses were also markedly different from other trophic morphs. As predicted, nectarivores were characterised by having generally narrow snouts, which is likely important in allowing the bat to penetrate the corolla of flowers (Koopman, 1981; Nicolay and Dumont, 2000) and to provide a stable support for the extremely long and mobile tongue (Koopman, 1981; Freeman, 1995). Based on the reduced importance of mastication in nectarivores, the jaw muscles need to be less developed, which would result in a low or narrow posterior part of the skull. However, we did not find a prominent reduction in the size of that area of the skull. Analysis of bite force in a nectarivorous species, however, did indicate reduced bite performance relative to other bats (Aguirre et al., 2002). One possibility is that the reduced bite performance is largely the result of an increase in the out-lever of the system. Alternatively, the inclusion of megachiropteran species in our group of nectarivores may have nullified any trend towards lower and narrower skulls. As indicated by Freeman's (1995) analysis of nectarivory in bats, megachiropteran nectarivores generally have wider and more robust skulls. This suggests that in these animals, nectarivory might not come at the cost of reduced bite performance. Bite force measurements in megachiropteran bats should provide further insights. Overall, our analyses indicate that evolutionary trends towards nectarivory are associated with a reduction in the width of the rostral part of the skull, but not with a marked reduction in head height or postorbital width. Sanguinivores were also very distinct from other trophic groups, as our analyses indicated significant differences in skull shape. As observed for carnivores, sanguinivores had higher and wider skulls than bats from other trophic groups, but in contrast to carnivores, sanguinivores also had narrow skulls at the level of the posterior upper molars. These results are partially in contrast to our predictions, as moderately long snouts, fairly narrow and low skulls were expected. These predictions were largely premised on the presumed decrease in performance due to the lack of mastication in these animals (see Aguirre et al. 2002). Clearly, reduced bite performance is not expressed by a reduction in skull height or width. As the lack of our understanding of the functional demands associated with this highly specialised feeding mode makes a priori predictions difficult, biomechanical and functional analysis of the jaw system in sanguinivores might be especially insightful. As a narrowing of the rostral part of the skull was also observed in nectarivorous bats, this trait is likely associated with the importance of the tongue during feeding. Frugivorous and insectivorous bat species also differed significantly in head shape. Generally, frugivores had wider rostral parts of the skull (i.e. large average values for Ior and c-c; note, however, that the width of the skull across the canines was no longer significant after taking into account the relationships among species), wider postorbital skulls and higher skulls than insectivores. Thus, skulls in frugivores were generally more robust than skulls in insectivores. This is likely related to the larger size of the food items typically consumed by frugivores as compared 216 Van Cakenberghe et al. to insectivores. Biting at larger gape angles in turn likely causes larger reaction forces on the skulls during forceful biting. As unilateral biting is a common behavioural strategy for frugivores (Dumont, 1999), this can potentially generate large torsional forces on the rostral part of the skull. The relatively wide rostra in frugivores might thus be a functional response to dissipate reaction forces generated while biting. In summary, the results indicate distinct morphological modifications associated with evolutionary shifts in diet. Although several of the changes in skull shape were in the predicted direction, other changes were unexpected. In contrast to the findings of Perez-Barberia and Gordon (1999) for ungulates none of the coronoid measurements were informative in discriminating among dietary groups. Skull height, on the other hand, did explain variation among dietary groups in both the bats examined here and among the ungulates examined by Perez-Barberia and Gordon (1999) and might thus be a generally informative character. Clearly, our understanding of the functional demands associated with some of these highly specialised feeding niches, and the biomechanics and functioning of the jaw system, is still insufficient to accurately predict evolutionary changes in skull shape associated with changes in diet. The results of the present study seem, at first sight, in contrast with recently published data in Aguirre et al. (2002). In that study, no correlations were observed between head shape and bite performance among phyllostomid bats. Although the head shape measurements included in that study are obviously an oversimplification of the true variation in head shape, one of the variables measured (head height) was significantly correlated with changes in diet in the present study. This suggests that some of the variation in skull shape observed here is not associated with bite force as a performance measure. Future studies examining bite performance in a wider range of bats including the families examined here will be useful in determining whether the absence of correlations between bite force and head shape is a general trait for bats. The results from this and previous studies (e.g. Perez-Barberia and Gordon, 1999) indicate that quantitative studies of head shape and diet in an explicit phylogenetic context are capable of detecting evolutionary trends, and are important in the general understanding of the evolution of dietary specialisation. Moreover our data indicate the importance of including a broad sample of radiations to detect general evolutionary trends. Future studies investigating bite performance and the biomechanics and function of the feeding system will be essential to reconstruct evolutionary pathways for dietary specialization in bats.

Acknowledgements

The authors which to thank the curators of the various museums who provided material for this study, especially Dr. Georges Lenglet of the Koninklijk Belgisch Instituut voor Natuurwetenschappen, Brussels, and Dr. Chris Smeenk of the Nationaal Natuurhistorisch Museum, Leiden. AH is a postdoctoral fellow of the Fund for Scientific research Flanders (FWO-Vl).

References

Acharya, L. (1992). Epomophorus wahlbergi. Mammalian Species, 394: 1 - 4, 3 figs. Aguirre, L.F., Herrel, A., Van Damme, R. and Matthysen, E. (2002) Ecomorphological analysis of trophic niche partitioning in a tropical savanna bat community. Proc. R. Soc. Lond. B. 269: 1271-1278. Alvarez, J., Willig, M.R., Jones, J.K. Jr. and Webster, W.D. (1991). Glossophaga soricina. Mammalian Species, 379: 1 - 7, 4 figs. Cranial Shape and Diet in Bats 217

Baeten, B., Van Cakenberghe, V. and De Vree, F. (1984). An annotated inventory of a collection of bats from Rwanda (Chiroptera). Rev. Zool. afr., 98 (1): 183 - 196. Baker, R.J. & Clark, C.L. (1987). Uroderma bilobatum. Mammalian Species, 279: 1 - 4, 4 figs. Barlow, K.E., Jones, G. and Barratt, E.M. (1997). Can skull morphology be used to predict ecological relationships between bat species? A test using two cryptic species of pipistrelle. Proc. R. Soc. Lond. B 264: 1695-1700. Bates, P.J.J. and Harrison, D.L. (1997). Bats of the Indian subcontinent. Harrison Zoological Museum Publication, Sevenoaks. 258 pp. Bernard, E. (2001a). Vertical stratification of bat communities in primary forests of Central Amazon, Brazil. J. Trop. Ecol., 17: 115 - 126. Bernard, E. (2001b). Species list of bats (Mammalia, Chiroptera) of Santarém area, Pará State, Brazil. Revta bras. Zool, 18 (2): 455 - 463. Bizerril, M.X.A. and Raw, A. (1998). Feeding behaviour of bats and the dispersal of Piper arboreum seeds in Brazil. J. Trop. Ecol., 14: 109 - 114. Brown, K.M. and Dunlop, J. (1997). Rhinolophus landeri. Mammalian Species, 567: 1 - 4, 4 figs. Churchill, S. (1998). Australian bats. Reed New Holland, Sydney, New South Wales, Australia: 230 pp. Claessen, C.J. and De Vree, F. (1990a). Systematic and distributional notes on the larger species of the genus Epomophorus Bennet, 1836 (Chiroptera: Pteropodidae). In: Vertebrates in the tropics. Museum Alexander Koenig, Bonn 1990 (ed. G. Peters, G. and R. Hutterer ), pp. 177 - 186. Claessen, C.J. and De Vree, F. (1990b). Systematic and taxonomic notes on the Epomophorus anurus-labiatus-minor complex with the description of a new species (Mammalia: Chiroptera: Pteropodidae). Senckenb. biol., 71 (4/6): 209 - 238. Cloutier, D. and Thomas, D.W. (1992). Carollia perspicillata. Mammalian Species, 417: 1 - 9, 3 figs. Corbet, C.B. and Hill, J.E. (1992). The mammals of the Indomalayan Region: A systematic review. Natural History Museum Publications - Oxford University Press: 488 pp. Cramer, M.J., Willig, M.R. and Jones, C. (2001). Trachops cirrhosus. Mammalian Species, 656: 1 - 6, 4 figs. De Faria, D.M. (1996). Food resource utilization by a Phyllostomidae phytophagous bat guild at the Santa Genebra Reserve, Campinas, SP, Brazil. Chiroptera Neotropical, 2 (1): 43. Dengis, C.A. (1996). Taphozous mauritianus. Mammalian Species, 522: 1 - 5, 4 figs. De Vree, F. (1969). Notes on Xiphonycteris spurrelli Dollman from the Togo Republic (Microchiroptera - Molossidae). Rev. Zool. Bot. afr., 80 (3-4): 273 - 280. De Vree, F. (1971). Notes sur une collection de Chiroptères ouest africains, principalement de la Côte d'Ivoire. Rev. Zool. Bot. afr., 83 (1-2): 31 - 49. De Vree, F., De Roo, A. and Verheyen, W.N. (1969). Contribution à l'étude des chiroptères de la République du Togo. Rev. Zool. Bot. afr., 80 (1-2): 200 - 207. De Vree, F., Hulselmans, J. and Verheyen, W.N. (1970). Contribution à l'étude des Chiroptères de la République du Togo. 2. Liste préliminaire des Chiroptères récoltés par la deuxième mission zoologique belge au Togo. Rev. Zool. Bot. afr., 82 (1-2): 41 - 46. Diaz-Uriarte, R. and Garland, T. Jr. (1998). Effects of branch lengths errors on the performance of phylogenetically independent contrasts. Syst. Biol. 47: 654-672. 218 Van Cakenberghe et al.

Dumont, E.R. (1999). The effect of food hardness on feeding behaviour in frugivorous bats (Phyllostomidae): an experimental study. J. Zool., Lond. 248: 219-229. Feldman, R., Whitaker, J.O. Jr. and Yom-Tov, Y. (2000). Dietary composition and habitat use in a desert insectivorous bat community in Israel. Acta Chiropt., 2 (1): 15 - 22. Felsenstein, J. (1985). Phylogenies and the comparative method. Am. Nat. 125: 1-15. Felten, H. (1956). Fledermäuse fressen Skorpione. Natur & Volk, 86: 53 - 57. Ferrell, C.S. and Wilson, D.E. (1991). Platyrrhinus helleri. Mammalian Species, 373: 1 - 5, 4 figs. Freeman, P.W. (1979). Specialised insectivory: beetle-eating and moth-eating molossid bats. J. Mamm. 60(3): 467-479. Freeman, P.W. (1981). A multivariate study of the family Molossidae (Mammalia: Chiroptera): morphology, ecology, evolution. Fieldiana Zool. 7: 1-173. Freeman, P.W. (1984). Functional cranial analysis of large animalivorous bats (Microchiroptera). Biol. J. Linn. Soc. 21: 387-408. Freeman, P.W. (1995). Nectarivorous feeding mechanisms in bats. Biol. J. Linn. Soc. 56: 439-463. Freeman, P. W. (2000). Macroevolution in Microchiroptera: recoupling morphology and ecology with phylogeny. Evol. Ecol. Res. 2: 317-335. Fujita, M.S. and Kunz, T.H. (1984). Pipistrellus subflavus. Mammalian Species, 228: 1 - 6, 3 figs. Gannon, M.R., Willig, M.R. and Jones, J.K. Jr. (1989). Sturnira lilium. Mammalian Species, 333: 1 - 5, 4 figs. Gardner, A.L. (1977). Feeding habits: In Biology of the bats of the New World family Phyllostomatidae, Part II. (ed. R.J. Baker, Jones, J.K. and Carter, D.C.). Spec. Publ. Mus. Texas Tech Univ., 13: 293 - 350. Garland, T.Jr., Dickerman, A.W., Janis, C.M. and Jones, J.A. (1993). Phylogenetic analysis of covariance by computer simulation. Syst. Biol. 42: 265-292. Harvey, P.H. and Pagel, M.D. (1991). The comparative method in evolutionary biology. Oxford: Oxford University Press. Happold, D.C.D. (1987). The mammals of Nigeria. Claredon Press - Oxford. 395 pp. Harrison, D.L. and Bates, P.J.J. (1991). The mammals of Arabia. Second Edition. Harrison Zoological Museum Publication, Sevenoaks: i - xv, 1 - 354. Herrera, L.G. and Martinez del Rio, C. (1998). Pollen digestion by new world bats: effects of processing time and feeding habits. Ecology, 79 (8): 2828 - 2838. Herrera M., L.G., Hobson, K.A., Manzo, A., Estrada B., D., Sánchez-Cordero, V. and Méndez C., G. (2001). The role of fruits and insects in the nutrition of frugivore bats: evaluating the use of stable isotope models. Biotropica, 33 (3): 520 - 528. Hood, C.S. and J. Pitocchelli (1983). Noctilio albiventer. Mammalian Species, 197: 1 - 5, 4 figs. Horácek, I., Hanák, V. and Gaisler, J. (2000). Bats of the Palearctic region: a taxonomic and biogeographic review. In Proceedings of the VIIIth EBRS. (ed. B.W. Woloszyn), pp. 11 - 157. Publication of CIC ISEZ PAN., Krakow Hutson, A.M., Mickleburgh, S.P. and Racey, P.A. (2001). Global status survey and conservation action plan: Microchiropteran bats. IUCN/SSC Chiroptera Specialist Group. Gland, Switzerland and Cambridge, U.K. i - x; 1 - 259. Cranial Shape and Diet in Bats 219

Ibáñez, C., Juste, J., Garcia-Mudarra, J.L. and Agirre-Mendi, P.T. (2001). Bat predation on nocturnally migrating birds. Proc. Natl. Acad. Sci. USA, 98 (14): 9700 - 9702. Jacobs, D.S. (1996). Morphological diversity in an insular bat, Lasiurus cinereus semotus. Funct. Ecol. 10(5): 622-630. Jones, K.E., Purvis, A., MacLarnon, A., Bininda-Edmonds, O.R.P. and Simmons, N.B. (2002). A Phylogenetic supertree of the bats (Mammalia: Chiroptera). Biol. Rev. 77: 223-259. Kalko, E.K.V. and Handley, C.O. Jr. (2001). Neotropical bats in the canopy: diversity, community structure, and implications for conservation. Plant Ecology, 153: 319 - 333. Kingdon, J. (1974). East African Mammals. An atlas of evolution in Africa. Volume 2. Part A. (Insectivores and Bats) Academic Press, London, New York. Kitchener, D.J., Gunnell, A. and Maharadatunkamsi (1990). Aspects of the feeding biology of fruit bats (Pteropodidae) on Lombok Island, Nusa Tenggara, Indonesia. Mammalia, 54 (4): 561 - 578. Koopman, K.F. (1981). The distributional patterns of New World nectar-feeding bats. Ann. Miss. Bot. Gard., 68: 352 - 369. Koopman, K.F. (1993). Chiroptera: In Mammal Species of the world. A taxonomic and geographic reference. Second Edition. (ed. D.E. Wilson and D.M. Reeder), pp. 137 - 241. Washington, London: Smithsonian Institution Press. Kwiecinski, G.G. and Griffiths, T.A. (1999). Rousettus egyptiacus. Mammalian Species, 611: 1 - 9, 3 figs. Langevin, P. and Barclay, M.R. (1990). Hypsignathus monstrosus. Mammalian Species, 357: 1 - 4, 4 figs. Lewis-Oritt, N., Van den Bussche, R.A. and Baker, R.J. (2001). Molecular evidence for evolution of piscivory in Noctilio (Chiroptera: Noctilionidae). J. Mamm., 82 (3): 748 - 759. Marshall, A.G. (1983). Bats, flowers, and fruit: evolutionary relationships in the Old World. Biol. J. Linn. Soc., 20: 115 - 135. Marshall, A.G. (1985). Old World phytophagous bats (Megachiroptera) and their food plants: a survey. Zool. J. Linn. Soc., 83: 351 - 369. McNab, B.K. and Bonaccorso, F.J. (2001). The metabolism of New Guinean pteropodid bats. J. Comp. Physiol., B 171: 201 - 214. Medellín, R.A. (1989). Chrotopterus auritus. Mammalian Species, 343: 1 - 5, 3 figs. Medellín, R.A. and Arita, H.R. (1989). Tonatia evotis and Tonatia silvicola. Mammalian Species, 334: 1 - 5, 4 figs. Nicolay, C.W. and Dumont, E.R. (2000). An experimental analysis of feeding performance in Syconycteris australis (Megachiroptera, Pteropodidae). Mammalia 64: 155-161. Perez-Barberia, F.J. and Gordon, I.J. (1999). The functional relationships between feeding type and jaw and cranial morphology in ungulates. Oecologia 118: 157-165. Peterson, R.L., Eger, J.L. and Mitchell, L. (1995). Faune de Madagascar: Chiroptères. Faune de Madagascar, 84: 1 - 204. Rice, W.R. (1989). Analysing tables of statistical tests. Evolution 43: 223-225. Robbins, C.B., De Vree, F. and Van Cakenberghe, V. (1985). Systematic revision of the African bat genus Scotophilus (Vespertilionidae). Ann. Kon. Mus. Mid. Afr., Zool. Wetensch., 246: 53 - 84. 220 Van Cakenberghe et al.

Rosevear, D.R. (1965). The bats of West Africa. Trustees of the British Museum (Nat. Hist.) London: i - xvii, 1 - 418. Rydell, J. (1993). Eptesicus nilssonii. Mammalian Species, 430: 1 - 7, 3 figs. Sazima, I. (1976). Observations on the feeding habits of phyllostomatid bats (Carollia, Anoura and Vampyrops) in southeastern Brazil. J. Mamm., 57 (2): 381 - 382. Simmons, N.B. and Voss, R.S. (1998). The mammals of Paracou, French Guiana: A Neotropical lowland rainforest fauna. Part 1. Bats. Bull. Am. Mus. Nat. Hist., 237: 1 - 219. Sinha, Y.P. and Advani, R. (1976). Notes on food and reproduction of some Rajasthan bats. Geobios, 3 (2): 37 - 40. Solmsen, E.-H. (1998). New World Nectar-feeding bats: Biology, morphology and craniometric approach to systematics. Bonn. zool. Monogr., 44: 1 - 118. Starrett, A. and de la Torre, L. (1964). Notes on a collection of bats from Central America, with the third record for Cyttarops alecto Thomas. Zoologica, 49: 53 - 63. Strait, S.G. (1993). Molar morphology and food texture among small-bodied insectivorous mammals. J. Mamm., 74 (2): 391 - 402. Turnbull, W.D. (1970). Mammalian masticatory apparatus. Fieldiana Geology 18: 149-356. Van Cakenberghe, V. and De Vree, F. (1985). Systematics of African Nycteris (Mammalia: Chiroptera). In Proceedings of the International Symposium on African Vertebrates Systematics, Phylogeny and Evolutionary Ecology, Bonn (ed. K.-L. Schuchmann), pp. 53 - 90. Bonn. Van Cakenberghe, V. and De Vree, F. (1993a). The systematic status of Southeast Asian Nycteris (Chiroptera: Nycteridae). Mammalia, 57 (2): 227 - 244. Van Cakenberghe, V. and De Vree, F. (1993b). Systematics of African Nycteris (Mammalia: Chiroptera). Part II. The Nycteris hispida group. Bonn. zool. Beitr., 44 (3/4): 399 - 332. Van Cakenberghe, V. and De Vree, F. (1994). A revision of the Rhinopomatidae Dobson 1872, with the description of a new subspecies (Mammalia: Chiroptera). Senckenb. biol., 73 (1/2): 1 - 24. Van Cakenberghe, V. and De Vree, F. (1999). Systematics of African Nycteris (Mammalia: Chiroptera) Part III. The Nyteris thebaica group. Bonn. zool. Beitr., 48 (2): 123 - 166 (for 1998). Van Cakenberghe, V., De Vree, F. and Leirs, H. (2000). On a collection of bats (Chiroptera) from Kikwit, Democratic Republic of the Congo. Mammalia, 63 (3): 291 - 322 (for 1999). Verschuren, J. (1957). Ecologie, biologie et systematique des Chiropteres. Explor. Parc Nat'l. Garamba, Miss. De Saeger (ed.). Inst. Parcs Natl. Congo Belge, Brussels 7: 1-473. Vonhof, M.J. and Kalcounis, M.C. (1999). Lavia frons. Mammalian Species, 614: 1 - 4, 3 figs. Wetterer, A.L., Rockman, M.V. and Simmons, N.B. (2000). Phylogeny of Phyllostomid bats (Mammalia: Chiroptera): Data from diverse morphological systems, sex chromosomes and restriction sites. Bull. Am. Mus.Nat. Hist., 248: 1 - 200. Whitaker, J.O. Jr. (1994). Food availability and opportunistic versus selective feeding in Insectivorous bats. Bat Research News, 34 (4):75 - 77. Whitaker, J.O. Jr. and Black, H. (1976). Food habits of cave bats from Zambia, Africa. J. Mamm., 57 (1): 199 - 204. Cranial Shape and Diet in Bats 221

Whitaker, J.O. Jr. and Mumford, R.E. (1978). Foods and ectoparasites of bats from Kenya, East Africa. J. Mamm., 59 (3): 632 - 634. Whitaker, J.O. Jr., Shalmon, B. and Kunz, T.H. (1994). Food and feeding habits of insectivorous bats from Israel. Z. Säugetierk., 59: 74 - 81. Whitaker, J.O. Jr. and Weeks, H.P. Jr. (2001). Food of Eptesicus fuscus, the Big Brown Bat, in Indiana in the absences of cultivated fields and agricultural pests. Proc. Indiana Acad. Sci., 110: 123 - 125. Willig, M.R., Camilo, G.R. and Noble, S.J. (1993). Dietary overlap in frugivorous and insectivorous bats from edaphic cerrado habitats of Brazil. J. Mamm., 74 (1): 117 - 128. Willis, K.B., Willig, M.R. and Jones, J.K. Jr. (1990). Vampyrodes caraccioli. Mammalian Species, 359: 1 - 4, 4 figs.

Appendix 1: Dietary classification of the species used in the present study (Ecomorph types: 1: Sanguinivore, 2: Nectarivore, 3: Carnivore, 4: Soft insect eaters, 5: Hard insect eaters, 6: Soft fruit eaters, 7: Hard fruit eaters, 8: Omnivores. N represents the number of specimens used in the analyses).

Species Diet Ecomorph N

Pteropodidae Megaloglossus woermanni * Palynivore, nectarivore (Rosevear, 1965); 2 26 flowers (Marshall, 1985) Melonycteris melanops * Nectarivore (Freeman, 1995) 2 1 Acerodon lucifer * Frugivore 6 1 Casinycteris argynnis * Frugivore 6 2 Cynopterus brachyotis * Frugivore (McNab and Bonaccorso, 2001); 7 3 flowers, fruit, leaves (Marshall, 1985) Dobsonia praedatrix * Frugivore (McNab and Bonaccorso, 2001); 6 1 Flowers and fruit (Marshall, 1985) Eidolon helvum Soft fruit, buds, young leaves, nectar, 7 3 pollen (Kingdon, 1974); Flowers, fruit and leaves (Marshall, 1985) Epomophorus anurus Frugivore 6 313 Epomophorus wahlbergi Fruit, leaves (Kitchener et al., 1990); mainly 7 214 figs, fruits of Diospyrus and Psidium, leaves of Balanites sp., sometimes beetles and other insects (Acharya, 1992). Epomops franqueti * Fruits, flowers (Happold, 1987) 6 2 Hypsignathus monstrosus * Figs, Anthocleista fruits, mangoes, bananas, 7 5 guavas (Happold, 1987); Flowers and fruit (Marshall, 1985); there are reports of it scavenging meat scraps and attacking chickens (Langevin and Barclay, 1990) Lissonycteris angolensis Frugivore (Marshall, 1985) 6 11 222 Van Cakenberghe et al.

Species Diet Ecomorph N Micropteropus pusillus * Bananas, figs, custard apples, flowers 7 69 (Rosevear, 1965); nectar (Happold, 1987) Nanonycteris veldkampi * Nectarivore (Rosevear, 1965); flowers 2 1 (Happold, 1987); Flowers and fruit (Marshall, 1985) Pteropus temmincki * Frugivore 6 2 Pteropus vampyrus * Frugivore 6 7 Rousettus aegyptiacus Soft fruits, buds, young leaves, nectar, 6 9 pollen (Kingdon, 1974); apple, apricot, banana, carob, date, fig, grape, jamba, litchis, loquat, mango, mulberry, orange, pawpaw, peach (Kwiecinski and Griffiths, 1999) Scotonycteris zenkeri Nectarivore (Freeman, 1995) 2 5 Rhinopomatidae Rhinopoma hardwickei Coleoptera, Isoptera, Lepidoptera (Sinha 5 182 and Advani, 1976) Rhinopoma macinnesi Insectivore 4 31 Rhinopoma microphyllum Coleoptera, Lepidoptera, Orthoptera, 5 77 Hymenoptera, Isoptera, Neuroptera, Dictyoptera (Bates and Harrison, 1997) Rhinopoma muscatellum Small insects (e.g. moths) (Bates and 4 75 Harrison, 1997) Emballonuridae Saccolaimus peli * Insectivore 4 2 Taphozous mauritianus Moths, termites, butterflies (Dengis, 1996) 4 17 Taphozous nudiventris Insectivore 4 27 Nycteridae Nycteris gambiensis Insectivore 4 231 Nycteris grandis Carnivore, including bats, geckos, mice 3 60 (Ibáñez et al., 2001); grasshoppers, moths, beetles, flies, small fish, amphibians, small birds (Happold, 1987) Nycteris hispida Small insects (e.g. moths) (Rosevear, 1965) 4 520 Nycteris javanica Insectivore 4 60 Nycteris macrotis Coleoptera, Orthoptera (Whitaker and 5 559 Black, 1976) Nycteris thebaica Scorpions, Coleoptera, Orthoptera, larval 5 572 Lepidoptera, Diptera, Hymenoptera, Hemiptera (Felten, 1956,Whitaker and Black, 1976; Feldman et al., 2000) Cranial Shape and Diet in Bats 223

Species Diet Ecomorph N

Megadermatidae Lavia frons Grasshoppers, beetles, mosquitoes, small 5 5 flies, small Vertebrates (Rosevear, 1965); solely soft and hardbodied insects > 2 mm (Vonhof and Kalcounis, 1999) Megaderma spasma Insectivore 4 1 Rhinolophidae Rhinolophus clivosus Coleoptera, Hymenoptera, Lepidoptera, 5 178 Trichoptera, Diptera, Homoptera, Neuroptera (Whitaker et al., 1994, Feldman et al., 2000) Rhinolophus ferrumequinum * Insectivore 4 10 Rhinolophus landeri Insectivore, primarily moths (Brown and 4 12 Dunlop, 1997) Hipposideridae Asellia tridens * Lepidoptera, Diptera, Hymenoptera, 4 10 Trichoptera, Coleoptera, Hemiptera, Homoptera, Neuroptera (Whitaker et al., 1994, Feldman et al., 2000) Cloeotis percivali * Lepidoptera (Whitaker and Black, 1976) 4 4 Hipposideros camerunensis Insectivore 4 3 Hipposideros commersoni Hard object insectivore (Strait, 1993); large 5 5 beetles (Happold, 1987) Hipposideros cyclops Insectivore 4 10 Hipposideros ruber Insectivore 4 49 Noctilionidae Noctilio albiventris * Insectivore (Lewis-Oritt et al., 2001); 5 7 Coleoptera, Hemiptera, Homoptera, Lepidoptera, Diptera, sometimes fish parts (Hood and Pitocchelli, 1983) Noctilio leporinus * Piscivore, Insectivore (Simmons and Voss, 3 9 1998; Lewis-Oritt et al., 2001) Phyllostomidae Chrotopterus auritus * Carnivore (Simmons and Voss, 1998; 3 3 Ibáñez et al., 2001; Kalko and Handley, 2001); Omnivore (Bernard, 2001a); small vertebrates, insects, fruit (Gardner, 1977); lizards, birds, shrews, mice, small bats, Insects (Ceramycidae, Scarabeidae, Sphingidae) (Medellín, 1989) Mimon crenulatum * Insectivore (Gardner, 1977; Bernard, 3 2 2001a; Kalko and Handley, 2001) 224 Van Cakenberghe et al.

Species Diet Ecomorph N Phyllostomus elongatus * Carnivore (Simmons and Voss, 1998); 8 12 Omnivore (Kalko and Handley, 2001); Insectivore (Bernard, 2001a) Phyllostomus hastatus * Omnivore (Willig et al., 1993, Simmons 8 10 and Voss, 1998; Bernard, 2001a; Kalko and Handley, 2001); insects, small vertebrates, and plant material, including fruit, pollen, nectar, and flower parts (Gardner, 1977) Tonatia brasiliense * Insectivore (Bernard, 2001a; Kalko and 4 2 Handley, 2001); gleaning animalivore (Bernard, 2001b); probably fruit and insects (Gardner, 1977) Tonatia schulzi * Carnivore (Simmons and Voss, 1998), 3 1 Insectivore (Bernard, 2001a) Tonatia silvicola * Insectivore (Bernard, 2001a; Kalko and 5 4 Handley, 2001); Coleoptera, Pedipalpida, Homoptera, Orthoptera, Hemiptera, fruit, Hymenoptera, whipscorpions (Medellín and Arita, 1989) Trachops cirrhosus * Carnivore (Simmons and Voss, 1998); 3 5 Insectivore (Bernard, 2001a; Kalko and Handley, 2001); gleaning animalivore (Bernard, 2001b); insects , small vertebrates and possibly some fruit (Gardner, 1977); opportunistic foliage- gleaning omnivore (Cramer et al., 2001). Vampyrum spectrum * Carnivore (e.g. bats, geckos, mice) 3 7 (Simmons and Voss, 1998; Ibáñez et al., 2001); birds, bats, rodents, and possibly some fruits and insects (Gardner, 1977) Anoura geoffroyi * Pollen, nectar, fruits, insects (Gardner, 2 6 1977); Insectivore , occasional pollen (Willig et al., 1993; Solmsen, 1998) nectarivore (Freeman, 1995) Glossophaga longirostris * Nectar, probably other flower parts, pulp 2 9 and fruit juice, occasional insects (Gardner, 1977; Solmsen, 1998); Nectarivore (Freeman, 1995) Glossophaga soricina * Frugivore (Bizerril and Raw, 1998); mainly 2 6 nectarivore (de Faria, 1996, Simmons and Voss, 1998; Kalko and Handley, 2001); nectar, flower parts, fruit, insects (Gardner, 1977; Solmsen, 1998); Omnivore (Bernard, 2001a); temporally and geographically variable diet (Alvarez et al., 1991); Nectarivore (Freeman, 1995) Cranial Shape and Diet in Bats 225

Species Diet Ecomorph N Lonchophylla thomasi * Nectarivore (Simmons and Voss, 1998; 2 2 Kalko and Handley, 2001); pollen, pulp and seeds of Piper sp., insects (Solmsen, 1998); Omnivore (Bernard, 2001a); Nectarivore (Freeman, 1995) Carollia perspicillata * Frugivore (de Faria, 1996; Simmons and 6 11 Voss, 1998; Bernard, 2001a; Kalko and Handley, 2001); largely frugivore but also catches insects (ca 2%), and laps nectar (Sazima, 1976; Willig et al., 1993); large variety of high protein and low fiber) fruits, some flowers, some (soft parts of hard-bodied) insects, complementary nectar and pollen (Cloutier and Thomas, 1992) Artibeus cinereus * Frugivore (Bernard, 2001a; Kalko and 6 2 Handley, 2001); fruits and insects (Gardner, 1977) Artibeus jamaicensis * Fruits, pollen, nectar, insects (Gardner, 7 12 1977); no pollen (Herrera and Martinez del Rio, 1998) fruits and pollen (Herrera et al., 2001); Frugivore (Bernard, 2001a; Kalko and Handley, 2001) Artibeus lituratus * Frugivore (de Faria, 1996; Bernard, 2001a; 7 7 Kalko and Handley, 2001); scattered insect remains (Starrett and de la Torre, 1964); nectar feeding (Sazima, 1976); insects and a variety of plant matter including fruit, flowers and leaves (Gardner, 1977) Artibeus obscurus * Frugivore (Bernard, 2001a; Kalko and 6 9 Handley, 2001) Platyrrhinus brachycephalus * Frugivore 6 4 Platyrrhinus helleri * Frugivore (Kalko and Handley, 2001); figs, 7 3 fruit of Acnistus, Cecropia sp. sometimes Lepidoptera (Ferrell and Wilson, 1991) Sphaeronycteris toxophyllum * Probably fruit (Gardner, 1977) 6 1 Sturnira lilium * Mainly frugivore (de Faria 1996; Bernard, 6 6 2001a); but also flowers, and presumably nectar and pollen (Gardner, 1977); no pollen (Herrera and Martinez del Rio, 1998), fruit, pollen, insects (Herrera et al., 2001); date palms, bananas, wild figs (Gannon et al., 1989) Uroderma bilobatum * Frugivore (Bernard, 2001a; Kalko and 7 8 Handley, 2001); various kinds of fruit and insects (Gardner, 1977); by volume 76 % plant material, 13 % insects and 11 % unclassified (Baker and Clark, 1987) 226 Van Cakenberghe et al.

Species Diet Ecomorph N Vampyrodes caraccioli * Frugivore (Gardner, 1977; Kalko and 6 1 Handley, 2001); figs, papaya, banana (Willis et al., 1990) Desmodus rotundus * Sanguivore (Bernard, 2001a; Kalko and 1 11 Handley, 2001) Diaemus youngii * Sanguivore (Bernard, 2001a; Kalko and 1 1 Handley, 2001) Vespertilionidae Arielulus societatis * Insectivore 4 8 Chalinolobus argentata Insectivore 4 3 Eptesicus bobrinskii * Insectivore 4 2 Eptesicus brasiliensis * Insectivore (Bernard, 2001a) 4 129 Eptesicus diminutus * Insectivore 4 12 Eptesicus floweri * Insectivore 4 2 Eptesicus fuscus * Hard object insectivore (Strait, 1993); 5 181 primarily Coleoptera (Withaker and Weeks, 2001) Eptesicus innoxius * Insectivore 4 8 Eptesicus nasutus * Insectivore 4 28 Eptesicus nilssoni * Small dipterans (3 - 10 mm), moths, beetles 5 32 (Rydell, 1993) Falsistrellus petersi * Insectivore 4 13 Falsistrellus tasmaniensis * Beetles, moths, bugs (Churchill, 1998) 5 19 Hypsugo arabicus * Insectivore 4 5 Hypsugo bodenheimeri * Lepidoptera, Diptera, Coleoptera 5 7 (Whitaker et al., 1994, Feldman et al., 2000) Hypsugo cadornae * Insectivore 4 9 Hypsugo eisentrauti * Insectivore 4 15 Hypsugo hesperus * Insectivore 4 74 Hypsugo imbricatus * Insectivore 4 4 Hypsugo kitcheneri * Insectivore 4 5 Hypsugo pulveratus * Insectivore 4 19 Hypsugo savii * Insectivore 4 53 Hypsugo stenopterus * Insectivore 4 11 Ia io * Insectivore 4 9 Myotis bocagei * Lepidoptera, Diptera, Coleoptera 5 11 (Whitaker et al., 1994) Myotis ridleyi * Insectivore 4 2 Neoromicia capensis * Beetles, moths (Happold, 1987) 5 133 Cranial Shape and Diet in Bats 227

Species Diet Ecomorph N Neoromicia rendalli * Lepidoptera, Hemiptera, Coleoptera 5 55 (Whitaker and Mumford, 1978) Neoromicia somalicus * Insectivore 4 75 Neoromicia tenuipinnis * Insectivore 4 55 Nyctalus joffrei* Insectivore 4 4 Nyctalus noctula * Insectivore 4 2 Nycticeius schlieffeni Beetles (Happold, 1987) 5 6 Perimyotis subflavus * Coleoptera, Homoptera, Diptera, 5 138 Hymenoptera, Lepidoptera, generally sized from 4 to 10 mm (Fujita and Kunz, 1984) Pipistrellus anchietae * Insectivore 4 11 Pipistrellus ceylonicus * Small beetles and other insects (Bates and 5 24 Harrison, 1997) Pipistrellus coromandra * Small flies, ants, Diptera (Bates and 4 20 Harrison, 1997) Pipistrellus flavescens * Insectivore 4 6 Pipistrellus javanicus * Insectivore 4 53 Pipistrellus kuhlii * Ants, Coleoptera, Chrysomelidae, 5 124 Lepidoptera, Cerambycidae (Whitaker et al., 1994, Feldman et al., 2000) Pipistrellus macrotis * Insectivore 4 17 Pipistrellus mimus * Beetles (Sinha and Advani, 1976) 5 40 Pipistrellus minahassae Insectivore 4 2 Pipistrellus nanulus * Insectivore 4 36 Pipistrellus nanus * Insectivore 4 109 Pipistrellus nathusii * Insectivore 4 49 Pipistrellus pipistrellus * Insectivore 4 54 Pipistrellus rueppelli * Lepidoptera, Trichoptera, Coleoptera, 5 37 Diptera (Whitaker et al., 1994; Feldman et al., 2000) Pipistrellus rusticus * Insectivore 4 51 Pipistrellus tenuis * Beetles, cockroaches, wingless ants, 5 53 grasshoppers, crickets, termites, moths, wasps, flies, mosquitoes (Bates and Harrison, 1997) Scotoecus hirundo Insectivore 4 7 Scotophilus borbonicus Insectivore 4 3 Scotophilus dinganii Coleoptera, Orthoptera, Hemiptera 5 310 (Whitaker and Mumford, 1978) Scotophilus leucogaster Insectivore 4 163 Scotophilus nigrita Insectivore 4 20 228 Van Cakenberghe et al.

Species Diet Ecomorph N Scotophilus nux Insectivore 4 34 Scotophilus viridis Coleoptera, Hemiptera (Whitaker and 5 144 Mumford, 1978) Scotozous dormeri * Beetles, moths, grasshoppers, crickets, 5 7 Neuroptera, termites, Orthoptera, Hymenoptera, Lepidoptera (Bates and Harrison, 1997) Vespadelus pumilus * Insectivore 4 28 Vespadelus regulus * Moths, flies, beetles, ants (Churchill, 1998) 4 16 Miniopterus inflatus Insectivore 4 41 Miniopterus schreibersi Mainly Lepidoptera, but also Coleoptera 5 10 (Whitaker and Black, 1976) Molossidae Chaerephon ansorgei * Formicidae, Hymenoptera (Verschuren, 4 102 1957) Chaerephon bemmeleni Insectivore 4 4 Chaerephon pumila * Hemiptera, Coleoptera (Whitaker et al., 5 156 1994) Mops condylurus * Coleoptera, Hemiptera (Whitaker et al., 5 8 1994) Mops nanulus * Insectivore 4 31 Mops niveiventer * Insectivore 4 34 Myopterus whitleyi * Insectivore 4 3 Otomops martiensseni * Soft insect eater (Freeman, 1979) 4 2 Tadarida fulminans Insectivore 4 10

Species indicated with an asterisk are those for which a full data set is available, and which were included in the initial principal component analysis. Cranial Shape and Diet in Bats 229

Appendix 2 Phylogenetic relationships among the species used in this study (based on Jones et al., 2002). Symbols indicate the trophic categories used to classify species. 230 Van Cakenberghe et al.

Pteropodidae Cranial Shape and Diet in Bats 231 232 Van Cakenberghe et al. Cranial Shape and Diet in Bats 233 234 Van Cakenberghe et al. Cranial Shape and Diet in Bats 235 236 Van Cakenberghe et al.